ABSTRACT. to discriminate precancerous tissue from other lesions by identifying ideal excitation wavelengths.

Transcription

1 GENDER MEDICINE/VOL. 9,NO. 1S, 2012 Oral Fluorescence Imaging Using 405-nm Excitation, Aiding the Discrimination of Cancers and Precancers by Identifying Changes in Collagen and Elastic Breakdown and Neovascularization in the Underlying Stroma Pierre Lane, PhD 1 ; Sylvia Lam, PhD 1 ; Michele Follen, MD, PhD 2 ; and Calum MacAulay, PhD 1 1 Department of Integrative Oncology, Section of Cancer Imaging, The British Columbia Cancer Research Centre, Vancouver, British Columbia, Canada; and 2 Department of Obstetrics and Gynecology, Drexel University of College of Medicine; Philadelphia, Pennsylvania ABSTRACT Background: Optical spectroscopy and imaging devices are being developed and tested for the screening and diagnosis of cancer and precancer in multiple organ sites. Objective: The aim of the study reported here is to optimize the capability of an optical imaging device to discriminate precancerous tissue from other lesions by identifying ideal excitation wavelengths. Methods: The studies reported here used a prototype of a direct fluorescence imaging device that uses 405-nm illumination to excite tissue. Results: There is ample evidence in the literature that 405 nm can distinguish oral cancers from normal tissue. Higher wavelengths may be necessary to differentiate potential confounding lesions, such as abrasions, burns, viral infections, inflammation, and gingivitis. Conclusions: Imaging at 405 nm could help doctors detect precancerous and cancerous oral lesions. Such imaging could be used by dentists, family practitioners, otorhinolaryngologists, general surgeons, obstetrician gynecologists, and internists, and could greatly increase the number of patients who have lesions detected in the precancerous phase. (Gend Med. 2012;9:S78 S82) 2012 Elsevier HS Journals, Inc. All rights reserved. Key words: cancer screening, fluorescence imaging, oral cancer diagnosis, oral intraepithelial neoplasia, sensitivity and specificity. Accepted for publication November 15, doi: /j.genm Elsevier HS Journals, Inc. All rights reserved /$ - see front matter S78

2 P. Lane et al. INTRODUCTION Almost all cancers are easier to treat successfully if caught early in their growth. This principal is particularly true for 85% of cancers that are epithelial in origin. Most of these cancers do not originate in fully malignant form. They typically arise from normal tissues that progress through one or more preinvasive pathologically recognizable steps, such as dysplasia or carcinoma in situ (CIS), for which there are limited detection methods. 1 Oral cancer is an especially tragic example of an epithelial cancer. 2 Although the mouth is easily accessible, oral cancers are often diagnosed at a late stage, beyond treatability, that brings with it high morbidity and mortality. Fluorescence spectroscopy and imaging may provide means of early detection and have been studied in many organ sites. 3 Robyler et al 4 reported on the use of 67 autofluoresence images that were taken of 56 patients who were divided into a training and validation set of 46 measurements each and applied to a test set of 21 images. An automated algorithm using 405-nm illumination discriminated normal from neoplastic tissue with a sensitivity of 96% and specificity of 96% in the training and in the validation sets and with a sensitivity of 100% and specificity of 91% in the test set. A previous report based on quantitative analyses suggested that imaging at this wavelength would be useful for diagnosis. 5 These results were confirmed by other investigators and found applicable to other organ sites This wavelength was shown to demonstrate intracellular activities of interest by several investigators This article presents the reader with a series of images showing the varied presentation of squamous cell carcinoma and dysplasia when observing these lesions with 405-nm excited direct fluorescence imaging, such as a clinician or dentist might see in an oral dysplasia referral clinic. We hope these images will be a useful visual aid as 405-nm excitation illumination fluorescence imaging devices are introduced into the market. MATERIALS AND METHODS Overview of Study Procedures Men and women who were 18 years of age were eligible to participate in the study, which began in 2008 and is ongoing. Patients are recruited from the oral cancer clinics of the British Columbia Cancer Research Centre and Agency. A research nurse or research assistant describes the study to eligible patients and written informed consent is obtained from those agreeing to participate, with emphasis that nonparticipation will not affect their care. A biomedical engineer assists with trial conduct, including assisting the provider with taking the photographic images included in the trial. All patients seen in the clinic receive: (1) a complete history and an oral cancer risk factors questionnaire, (2) a comprehensive head and neck examination, (3) a comprehensive oral white light examination, and (4) referrals for outstanding medical problems. Those patients in the trial undergo an additional examination using the device, which has violet light-excited fluorescence at 405 nm. Fluorescence images are taken of all lesions observed under white light or fluorescence. Suspicious areas are biopsied if the clinical impression suggest dysplasia or cancers. Each patient s history of lesions, including precancers and cancers are noted, and the clinical impression is entered into the database. Lesions that are biopsied are submitted for permanent section. All tissue biopsies are reviewed by histopathologists who are blinded to the imaging data and who specialize in oral cavity pathologic diagnosis. The study protocol was reviewed and approved by the Institutional Review Board at The British Columbia Cancer Agency. The studies were carried out at The British Columbia Cancer Agency, Vancouver, British Columbia, Canada. Fluorescence Imaging A handheld device emitting light at variable wavelengths was used. An image and detailed description of the imaging system can be found in Lane et al. 15 A light source with significant energy around 405 nm was positioned behind a highperformance, narrow band-pass interference filter designed to transmit predominately 405 nm light. This was used to excite the interior of the mouth, allowing the entire field of view of the camera to be illuminated. Each measurement takes approxi- S79

3 Gender Medicine mately an additional 3 minutes; however, documentation with photography adds an additional 2 minutes. A clinician using a similar device views the interrogated tissue through either filtered glasses or a variable magnification scope, both designed to highlight tissue fluorescence. Photographs presented here were taken using a filter to present the reader with an image akin to what a device operator sees. The total light exposure is less than the American National Standards Institute standards described in Roblyer et al. 4,5 All images were reviewed 4 times by 4 independent investigators blinded to the histology. These investigators included dentists, optical engineers, and clinicians familiar with fluorescence imaging. RESULTS One hundred nineteen patients participated in this study of 405 nm imaging. From each of these patients, several images were captured using fluorescence at 405 nm. Ninety percent of these patients had a concurrent biopsy at the time of their clinical visit; the remaining 10% were biopsied within 6 months of the time the image was taken. These images represent all surfaces areas in the oral cavity. Among the data set are 23 lesions of carcinoma (19 squamous and 2 verrucous); 2 CISs, 9 severe dyplasias, 18 moderate or mild dyplasias, 15 leukoplakias, 1 erythroplakia, 13 lichen planus, 5 hyperplasias, 2 mucositis or gingivitis, 10 traumatic lesions, 19 scars, 5 areas of pigmentation, including grafts and biopsy sites, and several other confounders (see Supplemental Appendix in the online version at doi: /j.genm ). For this article, we limited demonstrative images to invasive cancers, CIS, dyplasias, and erythroplakias and leukoplakias. The confounding lesions and scars will be published in a separate article outlining strategies for their separation from cancerous lesions. We chose 24 pairs of images (see Supplemental Figures 1 6 in the online version at doi: /j.genm ). On the left side, images are presented as seen by a clinician at the time of examination. On the right side, lesions are delineated to highlight regions of interest. Images of cancers, severe dysplasias, mild to moderate dysplasias, and 1 erythroplakia are presented. In most of these images, the delineated area is seen as a darker blue than the surrounding region. In some images, neovasculature can be noted. In those images where the delineated area is near the teeth, porphyrins can be noted, which fluoresce as pink, bright red, or pale green. There are several observations that can be made at 405 nm. The first observation is that areas of neoplasm that have outgrown their blood supply are often full of bacteria. Bacteria make porphyrins as a by-product, and porphyrins are seen as bright pink- and blue-colored areas using violet light fluorescence imaging. The dark areas demonstrate collagen and elastin breakdown in the stroma, and these areas can be very large, surrounding the white light visible cancers. Often, one can identify a plexus of vessels in the stroma that accompanies the breakdown of collagen and elastin. This neovascularization occurs as lesions grow; the process becomes increasingly deregulated in carcinogenesis. DISCUSSION Laser-induced fluorescence imaging has been used to identify precancerous and cancerous tissues for 20 years. When the field began, there were a limited number of lasers available: argon ion, helium neon, and helium cadmium. Besides being large and bulky instruments, they had limited lifetimes. Over the last 20 years, significant advances have made laser use more practical for clinical applications. Diode lasers at 405 nm have been used extensively in many forms of fluorescent microscopy. Their use has allowed cell biologists to image intracellular structures in the 20 to 50 nm range. They have also been used to turn on and off green fluorescent protein markers at the cellular level, facilitating a whole new field in science. This article shows a large number of images of 405 nm taken in a high-risk population in which there were a large percentage of dysplasic and neoplastic lesions. This gives the viewer some experience with images akin to the experience of the S80

4 P. Lane et al. imaging scientists and clinicians who performed the trials. It takes time to become accustomed to seeing the fluorescence images at this wavelength, and we provide a gallery of images for the viewer, to serve as a type of atlas. Most of the images show the lesion and an area of darkening surrounding it. Some of the images shown are quite dark in printed format. That is not so when viewed on the computer. This darkened image makes discrimination of the lesion harder when printed than when viewed on the computer. Early results suggested that the addition of the fluorescence examination at 405 nm was helpful in identifying characteristics indicative of precancer and cancer over that seen solely by comprehensive white light examination. Images of the violet-induced autofluorescence at 405 nm showed the deeper neovascularization of the stromal changes that accompanied the progression to neoplasia. Biologic studies of tissue slices demonstrated that there was loss of autofluorescence consistent with breakdown of collagen matrixes in the connective tissues as lesions progressed from normal to neoplastic. 3 Elastin breakdown is less well studied, but also strongly suspected in the carcinogenic pathway involving the epithelial stromal interface. Type 1 collagen is a major component of the extracellular matrix that exhibits both optical scattering and endogenous fluorescence. These optical properties render the study of the tumor microenvironment amenable to optical spectroscopy and imaging techniques. Further study of fluorescence imaging has the potential to improve clinical outcomes by giving providers more tools for the early detection of cancer. Similar technologies are being developed to employ fluorescence imaging in diverse organ sites, including the cervix, throat, and lungs. ACKNOWLEDGMENTS We would like to thank all the patients who selflessly participated in these studies in the interest of the advancement of medicine. This research would not have been possible without the contribution of Drs. Catherine Poh and Miriam Rosen of the British Columbia Cancer Research Agency and the University of British Columbia, who were responsible for collection and lesion demarcation shown in all of the image data. CONFLICTS OF INTEREST The symposium and publication of these proceedings were supported by: The National Institutes of Health, The Drexel University College of Medicine, The Helen I. Moorehead, MD Foundation, The Doris Willig, MD Foundation, The Institute for Women s Health and Leadership, and The Center for Women s Health Research at the Drexel University College of Medicine. Drs. Lane, Follen, and MacAulay and Lam have no financial interest in the device or company. This study was funded under a sponsored research agreement with Trimira, LLC. David A. Burns, Dr. Follen s husband, holds 8% interest in stock in Trimira. SUPPLEMENTAL MATERIAL Supplemental appendix and figures accompanying this article can be found in the online version at doi: /j.genm REFERENCES 1. World Health Organization. Accessed January 24, The Oral Cancer Foundation. The HPV connection. Accessed January 24, Richards-Kortum R, Sevick-Muraca. Quantitative optical spectroscopy for tissue diagnosis. Annu Rev Phys Chem. 1996;47: Roblyer D, Kurachi C, Stepanek V, et al. Objective detection and delineation of oral neoplasia using autofluorescence imaging. Cancer Prev Res (Phila). 2009;2: Roblyer D, Richards-Kortum R, Sokolov K, et al. Multispectral optical imaging device for in vivo detection of oral neoplasia. J Biomed Opt. 2008; 13: Da Costa RS, Andersson H, Wilson BC. Molecular fluorescence excitation-emission matrices rele- S81

5 Gender Medicine vant to tissue spectra. Photochem Photobiol. 2003;78: Andersson-Engels S, Elner A, Johansson J, et al. Clinical recording of laser-induced fluorescence spectra for evaluation of tumour demarcation feasibility in selected clinical specialties. Lasers Med Sci. 1991;6: Balasubramanian S, Elangovan V, Govindasamy S. Fluorescence spectroscopic identification of 7,12-dimethylbenz[a]anthracene-induced hamster buccal pouch carcinogenesis. Carcinogenesis. 1995;16: Farwell DG, Meier JD, Park Y, et al. Time-resolved fluorescence spectroscopy as a diagnostic technique of oral carcinoma: validation in the hamster buccal pouch model. Arch Otorhinolarygol Head Neck Surg. 2010;136: Lane P, Gilhuly T, Whitehead P, et al. Simple device for the direct visualization of oral-cavity tissue fluorescence. J Biomed Opt. 2006;11: Betzig E, Patterson GH, Sougrat R, et al. Imaging intracellular fluorescent proteins at nanometer resolution. Science. 2006;313: Melanson JE, Lucy CA. Violet (405) diode laser for laser induced fluorescence detection in capillary electrophoresis. The Royal Society of Chemistry. The Analyst. 2000;125: Beauregard KE, Lee K-D, Collier J, Swanson JA. ph-dependent perforation of macrophage phagoseomes by Listeriolysin O from Listeria moncytogenes. J Exp Med. 1997;186: Kirkpatrick ND, Hoying JB, Botting SK, et al. In vitro model for endogenous optical signatures of collagen. J Biomed Opt. 2006:11: Lane P, Follen M, MacAulay C. Has fluorescence spectroscopy come of age? A case series of oral precancers and cancers using white light, fluorescent light at 405 nm and reflected light at 545 nm using the Trimira Identafi Gender Med 2012;9:S25 S35. Address correspondence to: Michele Follen MD, PhD, Drexel University College of Medicine 245 N. 15th St., Philadelphia, PA Michele.Follen@Drexelmed.edu. S82

6 P. Lane et al. Supplemental Appendix 1. Study patients, date of measurement, location of lesions, and clinical impression Study ID Date Location Clinical Impression left lateral tongue 6mos scar for s/t of CIS right lateral tongue 1 yr old scar left ventral tongue 2 yr post s/t scar right lateral tongue 2.5 yr old scar, post s/t SCC right buccal mucosa 4 days wound, cheek biting lower left gingiva 6 mos post s/t graft right lateral tongue 6 mos post s/t scar for cancer left tongue background lichen planus right lateral tongue biopsy resident left buccal mucosa biteline right gingiva C1FV1 TB floor of mouth cancer, pre-surgery left tongue cancer, pre-surgery confounder left inside upper lip canker sore right buccal mucosa cheek biting right lower dentulus ridge CIS, concurrent bx left floor of mouth leukoplakia (D1-2) right posterior soft palate Squamous cell carcinoma right buccal mucosa confounder, contact lichenoid mucositis right lateral tongue mild dysplasia left lateral tongue moderate dysplasia right lateral tongue moderate dysplasia right lateral tongue moderate dysplasia right lateral tongue moderate dysplasia right lateral tongue moderate dysplasia right lateral tongue moderate to severe dysplasia left lateral tongue severe dysplasia floor of mouth C right tongue severe dysplasia left tongue severe dysplasia left lateral tongue dysplasia, hx of CIS right lateral tongue erythroplakia right lateral tongue fluorescence lost left lateral tongue geographic tongue tip of tongue geographic tongue left posterior tongue leukoplakia left ventral tongue leukoplakia left soft palate leukoplakia left ventral tongue leukoplakia right lateral tongue leukoplakia left tongue leukoplakia left lateral tongue leukoplakia left post maxillary retromolar leukoplakia right lateral tongue moderate dysplasia right buccal mucosa lichen planus right buccal mucosa lichen planus left posterior buccal mucosa lichen planus (continued) S82.e1

7 Gender Medicine Supplemental Appendix 1 (continued). Study ID Date Location Clinical Impression right lateral tongue lichen planus dorsal tongue lichen planus right buccal mucosa lichen planus left buccal mucosa lichen planus resident left gingival buccal lichen planus left buccal mucosa lichen planus, possible dysplasia right gingiva lichen planus, plaque type right buccal mucosa lichenoid dysplasia left lateral tongue mild dysplasia right lateral tongue mild dysplasia left lateral tongue mod dysplasia right lateral tongue mod dysplasia left lateral tongue mod dysplasia right lateral tongue mod dysplasia, leukoplakia left gingiva gingivitis right lateral tongue pigmentation left latetal tongue pigmentation right buccal mucosa pigmentation, lichen planus right ventral tongue post s/t scar right lateral tongue pre-surgery right tongue pre-surgery tongue pre-surgery right tongue pre-surgery left tongue pre-surgery left lateral tongue severe dyplasia right buccal mucosa red, white right tongue red, white right lateral tongue red, white leukoplakia right FOM residual lesion left ventral tongue scar left buccal mucosa scar right tongue scar right lateral tongue scar right lateral tongue scar left lateral tongue scar right lateral tongue scar right lateral tongue scar 3 mos rad tx severe dysplasia left lateral tongue scar, 6 mos post s/t CIS carcinoma in situ; hx history; SCC squamous cell carcinoma; s/t surgical treatment; rad tx radiotherapy. S82.e2

8 P. Lane et al. A B C D E F G H Supplemental Figure 1. Oral cavity illuminated at 405nm excitation. Areas of interest are delineated in panels on the right. S82.e3

9 Gender Medicine A B C D E F G H Supplemental Figure 2. Oral cavity illuminated at 405nm excitation. Areas of interest are delineated in panels on the right. S82.e4

10 P. Lane et al. A B C D E F G H Supplemental Figure 3. Oral cavity illuminated at 405nm excitation. Areas of interest are delineated in panels on the right. S82.e5

11 Gender Medicine A B C D E F G H Supplemental Figure 4. Oral cavity illuminated at 405nm excitation. Areas of interest are delineated in panels on the right. S82.e6

12 P. Lane et al. A B C D E F G H Supplemental Figure 5. Oral cavity illuminated at 405nm excitation. Areas of interest are delineated in panels on the right. S82.e7

13 Gender Medicine A B C D E F G H Supplemental Figure 6. Oral cavity illuminated at 405nm excitation. Areas of interest are delineated in panels on the right. S82.e8